A 64 Kbytes ISL - TAGE branch predictor ∗
نویسنده
چکیده
The ISL-TAGE predictor consists in a TAGE predictor combined with a loop predictor (to predict loops) , a Statistical Corrector predictor and an Immediate Update Mimicker, IUM. A TAGE predictor [4] constitutes the core of the ISL-TAGE predictor. The TAGE predictor (Section 2) captures most of the correlation on the branch outcomes for very long histories. But sometimes TAGE fails to predict loops with constant number of iterations. The loop predictor (Section 3 ) is used to predict these loops. TAGE also fails to predict branches that are not strongly correlated, but only statistically biased. The Statistical Corrector predictor (Section 4) is in in charge of tracking these statistically correlated branches that are not correctly predicted by the TAGE predictor. The TAGE predictor must be updated at retire time to avoid pollution by the wrong path; this delayed update induces extra mispredictions compared with an optimistic at fetch time update. The Immediate Update Mimicker (Section 5) aims at limiting these extra mispredictions through predicting branches with inflight non-retired occurrences. The parameters of the submitted ISL-TAGE predictor are summarized in Section 6 .
منابع مشابه
A 64-Kbytes ITTAGE indirect branch predictor∗
The ITTAGE, Indirect Target TAgged GEometric length predictor, was introduced in [5] at the same time as the TAGE conditional branch predictor. ITTAGE relies on the same principles as the TAGE predictor several predictor tables indexed through independent functions of the global branch/path history and the branch address. Like the TAGE predictor, ITTAGE uses (partially) tagged components as the...
متن کاملTAGE-SC-L Branch Predictors∗
The TAGE predictor [12] is considered as one of the most storage effective global branch/path history predictors. It has been shown that associated with small adjunct predictors like a statistical corrector (SC for short) and/or a loop predictor (L for short) [11, 10], TAGE can even be more effective. In this study, we explore the performance limits of these TAGE-SC-L predictors for respectivel...
متن کاملThe Second Journal of Instruction-Level Parallelism Championship Branch Prediction Competition (CBP-2)
The TAGE predictor, TAgged GEometric length predictor, was introduced in [10]. TAGE relies on several predictor tables indexed through independent functions of the global branch/path history and the branch address. The TAGE predictor uses (partially) tagged components as the PPM-like predictor [5]. It relies on (partial) match as the prediction computation function. TAGE also uses GEometric his...
متن کاملExploring branch predictability limits with the MTAGE+SC predictor∗
In the previous championship CBP-4, the winner of the unlimited storage track [5], poTAGE-SC was combining several TAGE based predictors using different forms of histories (local, global, and frequency), a COLT inspired [3] prediction combiner and a statistical corrector (SC) predictor [8, 10] fed with various forms of branch histories. With MTAGE-SC, we improve this predictor in two ways. Firs...
متن کامل